UNHCR Warns AI-Fuelled Misinformation Is Putting Refugees at Risk

Speaking at the AI for Good Summit in Geneva, UNHCR said artificial intelligence is reshaping the global information landscape in ways that bring both promise and danger.

The UN Refugee Agency (UNHCR) has warned that the rapid spread of misinformation, hate speech and AI-generated deepfakes is creating serious risks for refugees and humanitarian workers, making it harder for displaced people to access protection while increasing threats to aid operations around the world.

AI creates new challenges alongside new opportunities

Speaking at the AI for Good Summit in Geneva, UNHCR said artificial intelligence is reshaping the global information landscape in ways that bring both promise and danger. While AI can improve humanitarian services and help solve complex information challenges, it is also accelerating the spread of false information, manipulated content and online abuse.

The agency stressed that refugees and humanitarian organisations must not be overlooked in international discussions about AI governance. It called for stronger safeguards, better content moderation in humanitarian settings and improved support for less widely spoken languages, ensuring digital safety measures work effectively for vulnerable communities. UNHCR said technology companies, governments and researchers all have a role in creating AI systems that protect people rather than exposing them to greater harm.

False information is leading to real-world consequences

According to UNHCR, misinformation and hate speech are no longer limited to online platforms but are increasingly contributing to violence, discrimination and social division. False claims and dehumanising narratives have been linked to attacks on refugees, protests, restricted access to education and employment, and, in some cases, forced displacement and loss of life.

The agency highlighted recent events in Libya, where waves of online misinformation and hate speech fuelled hostility toward refugees while placing humanitarian workers at greater risk. UNHCR reported incidents involving AI-generated deepfake videos of its representatives, online posts revealing staff locations and messages encouraging violence against aid workers.

Communities such as the Rohingya, who have already fled persecution, continue to face widespread hate campaigns online even after seeking safety in other countries. UNHCR warned that these digital attacks deepen divisions between refugees and host communities while making integration more difficult.

UNHCR calls for stronger global cooperation

A recent UNHCR survey found that 93 percent of staff had witnessed misinformation, disinformation or hate speech affecting humanitarian operations. Women, including both refugees and humanitarian workers, were found to be disproportionately targeted, with generative AI making abusive content easier to produce and spread on a much larger scale.

The agency emphasised that protecting information integrity does not conflict with freedom of expression. Instead, it said the focus is on preventing harmful false information that can place lives at risk, including deceptive online messages from smugglers and traffickers who lure vulnerable people with fake promises of safe migration routes, jobs or legal opportunities.

To address these challenges, UNHCR has established a Community of Practice on Information Integrity in Humanitarian Contexts with support from Switzerland and continues to work with governments, technology companies and civil society. The agency also highlighted its Information Integrity Response Toolkit, developed with refugee communities and already used by around 2,000 humanitarian practitioners to assess risks and strengthen local responses.

UNHCR said maintaining trusted information has become an essential part of refugee protection and urged governments and technology companies to ensure humanitarian concerns remain central as global AI governance frameworks continue to evolve.